1. Field of the Invention
The present invention relates to a method of manufacturing an acceleration sensor having a fixed electrode and a movable electrode to be displaced according to an acceleration and serving to measure an acceleration based on a change in an electrostatic capacity between both electrodes.
2. Description of the Background Art
FIG. 13 is a sectional view showing an example of a structure of an acceleration sensor. The acceleration sensor comprises fixed electrodes 2a and 2b, a movable electrode 1, a surface side substrate 3, a back side substrate 4 and a frame portion 7.
The movable electrode 1 is displaced upon receipt of an acceleration. Accordingly, respective distances between the movable electrode 1 and the fixed electrodes 2a and 2b are changed. The change is detected as a variation in the electrostatic capacitance. Thus, the acceleration sensor converts the acceleration into an electric signal.
Moreover, FIGS. 14 and 15 are perspective views showing the acceleration sensor seen from the surface and back sides, respectively. FIG. 13 is a sectional view taken along the line A—A in FIGS. 14 and 15. In order to clearly illustrate an internal structure of the acceleration sensor, the back side substrate 4 and the surface side substrate 3 are separated from the acceleration sensor in FIGS. 14 and 15, respectively. Furthermore, FIG. 16 is a plan view showing the fixed electrodes 2a and 2b, the movable electrode 1 and the frame portion 7 in which the surface side substrate 3 is separated from the acceleration sensor.
As shown in FIGS. 13 to 16, the movable electrode 1 takes the shape of a rectangular parallelepiped having an H-shaped projection 1e on one surface. Then, beam-shaped portions 1a and 1c are extended from a set of opposite sides of the rectangular parallelepiped and the beam-shaped portions 1a and 1c are connected to movable electrode support portions 1b and 1d, respectively.
The projection 1e is provided on the movable electrode 1 for the following reason. The weight of the movable electrode 1 should be reduced to increase a sensitivity to the acceleration, and furthermore, the distance between the back side substrate 4 and the movable electrode 1 should not be greatly increased as compared with the distance between the surface side substrate 3 and the movable electrode 1.
The movable electrode support portions 1b and 1d are joined to the back side substrate 4 and the surface side substrate 3, and the movable electrode 1 is maintained in a floating state by the joined portions and the beam-shaped portions 1a and 1c. The back side substrate 4 and the surface side substrate 3 are provided with concave portions 4a and 3d such that the back side substrate 4 and the surface side substrate 3 do not come in contact with the movable electrode 1, respectively.
Moreover, the fixed electrodes 2a and 2b and the frame portion 7 are also joined to the back side substrate 4 and the surface side substrate 3. The movable electrode 1 is provided to be interposed between the two fixed electrodes 2a and 2b. The movable electrode 1 is not in contact with the two fixed electrodes 2a and 2b through a void 5.
The surface side substrate 3 is provided with contact holes 3a and 3b to be connected to the fixed electrodes 2a and 2b respectively and a contact hole 3c to be connected to the movable electrode support portion 1d. Electric potentials of the respective electrodes are detected from the contact holes 3a to 3c. Then, a change in an electrostatic capacity is detected between the electrodes. Only one fixed electrode can also be operated in the same manner. In that case, it is enough that the surface side substrate 3 should be provided with the contact hole 3a or 3b and the contact hole 3c. 
A method of manufacturing the acceleration sensor will be described below.
First of all, a silicon substrate for forming the movable electrode 1, the fixed electrodes 2a and 2b and the frame portion 7 and two glass substrates to be the surface side substrate 3 and the back side substrate 4 are prepared. The contact holes 3a to 3c and the concave portion 3d are formed in one glass substrate and the concave portion 4a is formed in another glass substrate. Thus, the surface side substrate 3 and the back side substrate 4 are formed.
In the silicon substrate, moreover, patterns of the movable electrode 1, the beam-shaped portions 1a and 1c, the movable electrode support portions 1b and 1d, the fixed electrodes 2a and 2b and the frame portion 7 are formed from the side of the surface side substrate 3 to the middle of a thickness of the substrate (corresponding to a thickness of the movable electrode 1) by using a photolithographic technique and an anisotropic etching technique.
Next, the silicon substrate and the surface side substrate 3 are joined to each other by using an anode junction method. As shown in FIG. 17, the silicon substrate 11 and the surface side substrate 3 are provided in alignment, and electrodes 13 and 15 are connected thereto and are heated, respectively. When a temperature suitable for the anode junction method is reached, a voltage is applied to the electrodes 13 and 15. At this time, a ground potential GND is applied to the electrode 13 connected to the silicon substrate 11, and an electric potential which is lower than the ground potential GND by an electric potential difference E is applied to the electrode 15 connected to the surface side substrate 3. In order to generate the electric potential difference E, it is preferable that a DC power supply 14 should be connected to the electrodes 13 and 15. By properly regulating the value of the electric potential difference E, a time required for application thereof, a temperature for junction and the like, a junction current is caused to flow to both electrodes so that the silicon substrate 11 and the surface side substrate 3 can be joined to each other.
Next, the voltage application is stopped and the electrodes 13 and 15 are removed from the silicon substrate 11 and the surface side substrate 3. By using the photolithographic technique and the anisotropic etching technique, the patterns of the projection 1e of the movable electrode 1, the movable electrode support portions 1b and 1d, the fixed electrodes 2a and 2b and the frame portion 7 are formed in a surface of the silicon substrate 11 which is opposite to the surface side substrate 3. Consequently, the movable electrode 1 is brought into a floating state.
Then, the anode junction method is used again to join the silicon substrate 11 to the back side substrate 4. As shown in FIG. 18, the silicon substrate 11 having the movable electrode 1, the fixed electrodes 2a and 2b and the frame portion 7 formed thereon and the back side substrate 4 are provided in alignment, and electrodes 13 and 15 are connected to the surface side substrate 3 and the back side substrate 4 and are heated, respectively. When a temperature suitable for the anode junction method is reached, a voltage is applied to the electrodes 13 and 15. At this time, a ground potential GND is applied to the electrode 13 connected to the surface side substrate 3, and an electric potential which is lower than the ground potential GND by the electric potential difference E is applied to the electrode 15 connected to the back side substrate 4. The same electric potential as that of the electrode 13 is applied to the frame portion 7 in the silicon substrate 11. By properly regulating the value of the electric potential difference E, a time required for application thereof, a temperature for junction and the like, a junction current is caused to flow to both electrodes so that the silicon substrate 11 and the back side substrate 4 can be joined to each other.
As shown in FIG. 19, then, a metal film 6 is formed over the whole surface of the surface side substrate 3 by using a sputtering method or the like. At this time, the metal film 6 is formed sufficiently thickly such that electrical connection can be attained between the metal film 6 and the movable electrode 1 and the fixed electrodes 2a and 2b in the contact holes 3a to 3c. 
As shown in FIG. 20, finally, the metal film 6 is subjected to patterning by using the photolithograpic technique and an etching technique and is divided into electrode films 6a, 6b and 6c connected to the fixed electrodes 2a and 2b and the movable electrode 1, respectively.
In the above-mentioned anode junction method, when the silicon substrate 11 and the back side substrate 4 are to be joined to each other, the movable electrode 1 and the fixed electrodes 2a and 2b are set in an electrical floating state because the movable electrode 1 and the fixed electrodes 2a and 2b are separated from the frame portion 7. For this reason, the movable electrode 1 and the fixed electrodes 2a and 2b are brought into an electrically unstable state so that an electric potential difference is easily made between the movable electrode 1 and the fixed electrodes 2a and 2b. In the anode junction method, usually, the value of the electric potential difference E is set to approximately 300 to 1000 V. In some cases, therefore, an electric potential difference of approximately several tens V is made between the movable electrode 1 and the fixed electrodes 2a and 2b. 
A distance between the movable electrode 1 and the fixed electrodes 2a and 2b is designed to be approximately several μm. For this reason, when the electric potential difference of approximately several tens V is made between the movable electrode 1 and the fixed electrodes 2a and 2b, Coulomb force (electrostatic attractive force) for attraction is generated so that both electrodes are bonded to each other. Such a phenomenon deteriorates the yield of the acceleration sensor so that the cost of the acceleration sensor cannot be reduced.